churchill-lab / emase

Expectation-Maximization algorithm for Allele-Specific Expression
http://churchill-lab.github.io/emase/
GNU General Public License v3.0
22 stars 13 forks source link

dealing with paired-end reads #7

Closed lassancejm closed 8 years ago

lassancejm commented 8 years ago

Hi,

Going through the examples, I noticed that the alignment step is always done considering single-end data. Any particular reason for that?

If not, do you think it is ok to use the following parameters for the alignment:

bowtie -a --best --strata -S -X 750 --chunkmbs 256 -p 16 ${bowtieidx} -1 ${prefix}.R1.fq -2 ${prefix}.R2.fq | samtools view -F 0xC -bS - >${prefix}_emase.bam

Thanks a lot!

kbchoi-jax commented 8 years ago

What I recommend for paired-end reads is to (1) run bowtie for each end independently, (2) convert each bam file to emase format, (3) combine two ends using the script, get-common-alignments, and then (4) run EM.

lassancejm commented 8 years ago

Will do as you recommend then, thank you.